Multi-scale composite reliability-based design optimization for aeroelastic applications

Ludovic Coelho (1) , Nicolò Fabbiane (1) , Christian Fagiano (1) , Cédric Julien (1) , Roger Ballester (1) , Didier Lucor (2, 3)

1 DAAA, ONERA, Université Paris-Saclay [Châtillon]
2 LISN - Laboratoire Interdisciplinaire des Sciences du Numérique
3 DATAFLOT - DAta science, TrAnsition, Fluid instabiLity, contrOl, Turbulence

The inherent incertitudes linked to the composite manufacturing process (e.g., material properties, thickness, ply orientations) call for a reliability-based approach to composite structural optimization [1]. Uncertainty is often introduced at lower scales of the composite material, while a more macroscopic scale is the preferred design space for optimization. This work proposes a new iterative methodology that combines a lowdimensional macroscopic design space-relying on homogenized variables-with gradient information to perform quick and accurate optimization and a high-dimensional lower-scale (ply scale) space where uncertain design variables are modeled and upscaled. An inverse problem is solved at each iteration of the optimization process to identify the low-scale material configuration that meets the homogenized parameters in terms of some statistical description. This resolution becomes viable thanks to efficient metamodel upscaling: a particular orthonormal basis is constructed with Fourier chaos expansion, which provides a very efficient closed-form expression of the macroscopic design variable statistics. Moreover, due to the multimodal nature of the aeroelastic response (i.e., the flutter velocity), a surrogate model strategy based on Gaussian process classification is implemented for the reliability analysis. This approach is applied to a composite plate optimization with uncertain ply angles to promote the wing's flexibility while remaining reliable with respect to the flutter phenomenon. The results show a good convergence of this optimization approach with a significant improvement in reliability compared to the deterministic optimized design (cf. Fig. 1) and a significant computational gain compared to the approach of directly optimizing ply angles via an evolutionary algorithm as in [2]. Optimization path of the Reliability-Based Design Optimization. Final macroscopic design variables are compared with deterministic optimization.

Domaines: Mécanique des fluides [physics.class-ph]; Apprentissage [cs.LG]; Matériaux et structures en mécanique [physics.class-ph]

Ludovic Coelho, Nicolò Fabbiane, Christian Fagiano, Cédric Julien, Roger Ballester, et al.. Multi-scale composite reliability-based design optimization for aeroelastic applications. 5th International Conference on Uncertainty Quantification in Computational Science and Engineering, Jun 2023, Athens, Greece. ⟨hal-04463192⟩